Determinants of the Quality of Warfarin Control after Venous Thromboembolism and Validation of the SAMe-TT2-R2 Score: An Analysis of Hokusai-VTE

Stefano Barco, Serena Granziera, Michiel Coppens, Jonathan Douxfils, Mathilde Nijkeuter, Nicoletta Riva, Thomas Vanassche, George Zhang, Min Lin, Pieter W Kamphuisen, Alexander T Cohen, Jan Beyer-Westendorf

Research output: Contribution to journalArticlepeer-review

6 Downloads (Pure)


Background Time in therapeutic range (TTR) measures the quality of vitamin K antagonist (VKA) anticoagulation. In patients with atrial fibrillation, the dichotomized SAMe-TT2-R2 score (≥2 vs. < 2 points) can predict if adequate TTR is unlikely to be achieved. Aims We validated the SAMe-TT2-R2 score in patients with venous thromboembolism (VTE) randomized to the warfarin arm of the Hokusai-VTE trial. Patients and Methods A total of 3,874 patients were included in the primary analysis (day 31-180 from randomization). The efficacy and safety outcomes were symptomatic recurrent VTE and major or clinically relevant non-major bleeding. Results The rates of recurrent VTE and bleeding events were higher in patients with a TTR below the median (< 66% vs. ≥66%) resulting in an absolute risk difference (ARD) of +0.5% (95% confidence interval: 0%, +1.1%) and +2.2% (0.9%, +3.5%), respectively. Patients with high SAMe-TT2-R2 score were 76% of total and had lower median TTR (64.7% vs. 70.7%). The SAMe-TT2-R2 score exhibited low negative (0.59) and positive (0.52) predictive value (TTR threshold 66%), and poor discrimination (c-statistic, 0.58). ARD between patients with high and low score was 0% (-0.6%, +0.7%) for recurrence and +1.3% (-0.1%, +2.7%) for bleeding. Results were confirmed in sensitivity analyses focusing on the whole study period (day 1-365). Conclusion In VTE patients, the SAMe-TT2-R2 score showed unsatisfactory discrimination and predictive value for individual TTR and did not correlate well with clinical outcomes. The choice of starting a patient on VKA cannot be based on this parameter and its routine use after VTE may not translate into clinical usefulness.

Original languageEnglish
Pages (from-to)675-684
Number of pages10
JournalThrombosis and Haemostasis
Issue number4
Publication statusPublished - 1 Jan 2019


  • quality of treatment
  • risk assessment model
  • venous thromboembolism
  • vitamin K antagonist
  • warfarin

Cite this